karpathy-guidelines

name	karpathy-guidelines
description	Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.
license	MIT

Karpathy Guidelines

Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Tradeoff: These guidelines bias toward caution over speed. For trivial tasks, use judgment.

1. Think Before Coding

Don't assume. Don't hide confusion. Surface tradeoffs.

Before implementing:

• State your assumptions explicitly. If uncertain, ask.
• If multiple interpretations exist, present them - don't pick silently.
• If a simpler approach exists, say so. Push back when warranted.
• If something is unclear, stop. Name what's confusing. Ask.

2. Surface Unknowns Before Committing

The map (prompts, skills, context you were given) is not the territory (codebase, real constraints, actual users). Before implementing, classify the task's unknowns — the quadrant drives the move:

• Known knowns (in the prompt) — trust but verify; stale ones are worse than admitted gaps.
• Known unknowns (you know the gap exists) — map only what this decision needs, not everything.
• Unknown knowns ("I'd know it on sight": taste, conventions, fit) — surface them with a prototype or a few options to react to. Finding these during implementation is expensive: a small spec change can force a different structure and a costly revert.
• Unknown unknowns (you don't know what you don't know) — ask, up front, what questions you don't yet know to ask and what "good" even looks like.

When unknown-knowns or unknown-unknowns dominate, prefer a cheap probe (brainstorm, throwaway prototype, one-question-at-a-time interview) over committing to implementation. Discovery is cheap; reverts are not. If unknowns are high, ask the user for their starting point and experience first.

3. Simplicity First

Minimum code that solves the problem. Nothing speculative.

• No features beyond what was asked.
• No abstractions for single-use code.
• No "flexibility" or "configurability" that wasn't requested.
• No error handling for impossible scenarios.
• If you write 200 lines and it could be 50, rewrite it.

Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.

4. Surgical Changes

Touch only what you must. Clean up only your own mess.

When editing existing code:

• Don't "improve" adjacent code, comments, or formatting.
• Don't refactor things that aren't broken.
• Match existing style, even if you'd do it differently.
• If you notice unrelated dead code, mention it - don't delete it.

When your changes create orphans:

• Remove imports/variables/functions that YOUR changes made unused.
• Don't remove pre-existing dead code unless asked.

The test: Every changed line should trace directly to the user's request.

5. Goal-Driven Execution

Define success criteria. Loop until verified.

Transform tasks into verifiable goals:

• "Add validation" → "Write tests for invalid inputs, then make them pass"
• "Fix the bug" → "Write a test that reproduces it, then make it pass"
• "Refactor X" → "Ensure tests pass before and after"

For multi-step tasks, state a brief plan:

1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]

Strong success criteria let you loop independently. Weak criteria ("make it work") require constant clarification.